Chatbots and voice consultants have firmly entered the service sector. Although I thought a few decades ago that the electronic voice that solves your problems is a story from a fantasy book. These e-advisers are getting smarter with advances in computing technology and AI training methods. I give this example as the most clear and understandable to any person. Now back to manufacturing.

The main difference between the sphere of manufacturing and, for example, the sphere of trade is costs. So in the mass market, logistics, advertising and customer service occupy the main share of all costs. In manufacturing (we are talking about large-scale manufacturing), the main costs are raw materials and equipment. Consequently, the following tasks arise that reduce costs: to use raw materials rationally, to purchase materials of optimal quality, to extend the life of equipment. Artificial intelligence is able to cope with these tasks.

Irene, please give examples of the use of artificial intelligence in manufacturing.

For example, a film production machine selects the optimal temperature regime depending on the specified output criteria (film thickness, environmental resistance) and the viscosity of the solution. This rationalizes the use of raw materials and energy resources. A large production machine has vibrations. According to the characteristic peaks of the curves that correspond to these vibrations, the engineer receives a message about the replacement of a particular part. Timely repair of the unit helps to avoid costly repairs. And it is often simply impossible to manually check the production machine due to forced downtime.

What are the most common challenges to making manufacturing smarter?

I highlight two issues. This is management and data. Often the management does not want or does not understand why they need to introduce technologies in manufacturing. It seems to them that there are more significant problems, and this will have to wait. But it is still waiting until more modern competitors are forced out of the market or the costs of manufactoring are higher than income.

The second problem, as I said, concerns the set of data. Everything is simple here. Often it just don’t exist. In order to implement technologies based on models, these models must be created. At the heart of everything is the accumulated experience, which can be digitized. That is, in addition to interviewing personnel, a system of sensors is also needed that will capture the necessary information.

Will manufacturing be smart or is it a failed strategy?

They are already smart and will get smarter in the future. I am optimistic. As practice shows, everything rests on the development of technology. Old technologies are being replaced by new ones. Technology is becoming cheaper and therefore more accessible. Yesterday there was a floppy disk, and today it’s an SSD disk. The development of wireless communications, the Internet of things, and so on contributes to the creation of smart manufacturing. The main task of a consultant in the field of digitalization of manufacturing today is to convey the need for this entire process to manufacturing management.

by Risher Panir,

CounselSenses

The post Expert opinion about Future of Smart Manufacturing appeared first on strategAI.

The system was developed in 1998 by Sergey Brin and Larry Page, graduates of Stanford University, who used PageRank technology to rank documents. In this technology, one of the key points is to determine the “authority” of a particular document based on information about the documents that refer to it. Along with this, Google used to determine the relevance of a document not only the text of the document itself, but also the text of links to it. This technology has allowed him to provide quite relevant results compared to other search engines. Quite quickly, Google began to lead in various surveys on such an indicator as user satisfaction with search results.

Google searches documents in over 35 languages. Nowadays, many portals and specialized sites provide Google-based information search services on the Internet, which makes the task of successfully positioning sites on Google even more important. Google re-indexes its search database approximately every four weeks. During this process (unofficially called Google dance), the database is updated based on information collected by robots and the PageRank values ​​of documents are recalculated. There is also a certain number of documents with a fairly high PageRank value, information about which in the search base is updated daily.

There are three main processes that allow us to get the expected result: crawling, indexing and search engine results.

Crawling is the process by which Googlebot discovers new and updated pages to add to the Google index. This process uses a large number of computers to efficiently scan the content of many web pages. The program that is responsible for performing the crawling is called a Google robot (or another common name, spider). The spider’s algorithm has the following points: programs determine which sites to crawl, how often to do it, and how many pages should be selected on each site. Google primarily starts with a list of web page URLs that was generated during previous crawls, augmented by data from sitemaps.

Sitemap (sitemap) are XML files with information for search engines, which is aimed at helping search engines to determine the location of the pages of the site, when they were last updated, the frequency of updates and the importance in relation to other pages on the site.

The Googlebot search robot finds links on each page and supplements the list of pages to be crawled, marks new and broken links. Robots send data about web pages to servers.

Indexing is the processing of a page in order to compile a complete index of the found words, as well as mark in which part of the page they are located. Moreover, the robot processes data from the main tags and attributes. However, some multimedia files and dynamically generated pages cannot be processed.

The search engine results are a response to a custom query. When a user enters a query, Google performs a search in the database that matches the terms and algorithmically determines the relevance of the content. After entering a search query, the system finds the corresponding pages in the index and the user receives the most relevant results.

Relevance refers to the degree to which search results meet user expectations. By this degree of issue, one can judge the effectiveness of the search engine. Relevance is determined by various factors, of which there are more than 200. For example, it can be the factor of incoming links from other pages. Each link to a page from another site increases the PageRank of the resource. To ensure that your site is ranked adequately on search engine results pages, it is important that the Googlebot can crawl and index your site correctly. It should be noted that for an objective assessment of the significance of a web resource, PageRank takes into account more than 10⁸ variables and 10⁹ terms.

Moreover, over 200 criteria are applied for processing requests. Google makes about 500 improvements to its search algorithm every year.

by Mason Flores

The post How to Works Google Search Machine appeared first on strategAI.

In 2003, Moore admitted that exponential growth was not possible in the future due to technological constraints. In 2007, he announced that the law would end soon. The number of transistors on an IC is mainly changed by decreasing their size. However, the reduction of the classical transistor leads to undesirable effects that are caused by the dimensional limits. In addition, the IC manufacturing technology also has its limitations. Device designs have to be constantly improved or breakthrough technologies have to be offered. Scientists and engineers are working on the creation of new designs, but the limit of semiconductor technology will be reached. Fundamentally new approaches are needed. For example, spin transistors are investigated, where the principle of operation is shifted from an electron charge to magnetic moment.

One of the main conclusion from this law is with an increase in processor power, the cost of computing equipment falls. Over the past 50 years, the cost of manufacturing a transistor has dropped thousands of times. Computers are now available in every home. Hence our time is called the digital era. We cannot imagine modern business without software for automating business processes. And all traditional marketing has been replaced with digital. AI, big date, remote work, all of this appeared only thanks to the development of electronics.

Moore himself refuted the work of his law in the future. Does this mean that we have reached the limit? There are many development options. For example, new materials (2D) are being actively studied by researchers around the world. Two-dimensional materials are not much more than 10 years old, and there are already a large number of solutions for electronic components based on this. New technologies and new fundamental principles of work will help to overcome this limit for sure.

¹Gordon Moore is one of the founders of Intel.

²The transistor is the main component of any electronic device. It is used for generating, switching and converting signals. Most often the transistor is made from semiconductors (semiconductor technology).

³Eugene Meieran is Intel engineer.

by Maria Baranova

The post How the Development of Radio Electronics Affect the Whole World Economy appeared first on strategAI.

Machine learning is the study of computer algorithm that have ability to automatically learn and improve from experience. Models are improved (become more accurate) due to the accumulated data. Thus, AI using machine learning is able to recognize patterns in data and generate predictions or reactions.

What are the benefits of AI for business:

1. Make predictions (rationale for business strategy),
2. Business process automation,
3. Quality control of work,
4. Marketing.

For example, personalization is a major trend in modern business. This is actively used by social media. Another example is dynamic pricing, which is actively used by low-cost airlines. Some brands have launched applications where you can try on their products online using augmented reality and sophisticated image recognition algorithms. For example, sneakers from a well-known corporation can be tried on by pointing the camera at your feet.

What caused the rapid development of AI

The development of AI is closely related to the development of computing. The computing power and memory required are large for machine learning and big data. The first semiconductor transistors were developed 50 years ago and were centimeter-sized. Six transistors are needed to write one bit of information. How many such transistors are needed for a complex calculation that writes gigabytes of information? Modern transistors are about 10 nanometers in size. Over 50,000 transistors can be placed along the hair diameter. Megabytes of memory can be placed on a hair slice. Thus, the evolution of technology (in particular micro and nanoelectronics) has led to the evolution of AI.

Who can implement AI in their business

The market is filled with ready-made solutions. You don’t have to have big budgets and develop your own solution. AI should be implemented with a clear understanding of what you need. First of all, innovations are introduced in the area that will bring the greatest benefit. AI will not replace employees, but it will help them.

by Maria Baranova

The post Basic Knowledge of Artificial Intelligence in Business Sphere appeared first on strategAI.

The volume of gas consumption is growing every year. Demand is expected to increase by 45% by 2030. The demand for gas is conditioned not only on the demand by the power industry, but also on the use of this fuel as a motor fuel. For example, the USA and Russia have developed a program and even adopted a maritime convention for the conversion of sea-going vessels to gas fuel. Limiting emissions in the sea also stimulates the transition to greener options. For example, from 2023, a limit on engine emissions has been established in the Baltic Sea. Also, the demand driver is the global trend for eco-friendly cars.

Delivery of gas to consumers is becoming an urgent task for gas producing companies. There are two ways to solve this problem.

• The first is delivery by means of a gas pipeline. In this case, there must be favorable geographic conditions, that is an economically advantageous mutual location of the gas deposits and its consumers. Mostly Russian deposits correspond to such conditions. Russia has a unique geographic location with the main consumers from all majpr gas consuming parts of the world near-by (Asia and Europe).

• The second way is gas liquefaction and transportation by gas carriers. In this case, it is necessary to design liquefaction production facilities for the deposit. Simply put – factories. At the design stage of the plant, it is necessary to understand what technology will be used for liquefaction. Equipment suppliers will be selected depending on the technology.

LNG technology is a complex process that varies greatly from liquefaction volumes. There are factories with a small liquefaction volume of up to 1.5 million tons per year, there are factories with large liquefaction volumes (10 million or more). Today only 4 companies have unique liquefaction technologies: Air Products (USA), Shell (Netherlands and Britain), Linde AG (Germany), ConocoFilips (USA). However, this list will be updated with new names. For example, the Russian company Novatek is currently developing its own liquefaction technology. This company will be able to take a leading place among others, subject to the success of their development.

The LNG industry, in addition to building a plant with a specific liquefaction technology, also needs the production of gas carriers. We will not go into details of this industry, but note that this area requires a complex technological process. This is due to the need to control and comply with the specified thermodynamic parameters during transportation.

Liquefied natural gas export volumes have recently increased by 30%. The main consumer countries are Japan, China, South Korea, Taiwan, India. According to data for 2016, Japan consumes somewhere in the order of 30%. This country is an island state, so the issue of transportation is very acute for them. The main LNG producers are Qatar, Australia, Malaysia, Algeria, USA, Nigeria, Russia. Many of these leaders, due to their geographic location, cannot provide pipeline gas. Therefore, both from the point of view of the supplier and from the point of view of the consumer, the relevance of LNG is beyond doubt.

The LNG market will only grow for at least the next 10 years. During this time, no source of alternative energy will be able to meet the demand, despite the diligence of the scientific community. There are a lot of gas reserves on the planet. Many deposits have not yet been explored and developed. Now a fierce struggle is being waged for the Arctic, since about 30% of the world’s gas reserves are located on its shelf. The explored gas fields will last for the next 150-200 years.

by Oleg Yakimov

The post Liquefied Natural Gas Market Prospects appeared first on strategAI.